Finds the include path for probematic files in a DAG from CL output

cl_include_tree_search.py

import sys
from pathlib import Path
from functools import partial, reduce
from pprint import pprint

eval_path = partial(reduce, lambda d, k: d[k])

def extract_includes(output_path):
    results = []
    with open(output_path, 'r', encoding='utf-16') as f:
        while True:
            line = f.readline()
            if not line:
                break

            _, sep, include = line.partition("Note: including file:")
            if include:
                results.append(include)
    return results

def build_tree(includes):
    tree = dict()
    cur_path = []

    def level(inc_line):
        return (len(inc_line) - len(inc_line.lstrip(' '))) -1 # there is a leading space for all

    for i in includes:
        lvl = level(i)
        if lvl > len(cur_path):
            pass # We don't need to back out anything on the path
        elif lvl == len(cur_path):
            if cur_path: # handle root sibling case
                cur_path.pop()
        elif lvl < len(cur_path):
            cur_path = cur_path[:lvl - 1]

        key = i.strip()
        eval_path(cur_path, tree).setdefault(key, {})
        cur_path.append(key)
            
    return tree


test = """\
  level 1
  level 2
   level 2.1
    level 2.1.1
  level 3
   level 3.1
   level 3.2
    level 3.2.1""".splitlines()

def dfs(tree, pattern):
    results = []
    _dfs(tree, pattern, [], results)
    return results

def _dfs(tree, pattern, path, results):
    if path and pattern in path[-1]:
        results.append(path)
    
    for k in eval_path(path, tree).keys():
        _dfs(tree, pattern, [*path, k], results)

if __name__ == "__main__":
    cl_output = Path(sys.argv[1])
    includes = extract_includes(cl_output)

    #ttree = build_tree(test)
    itree = build_tree(includes)

    pprint(dfs(itree, 'dtype'))
    pprint(dfs(itree, 'system_category_win32'))